Re: [PR] Make FSTPostingFormat to build FST off-heap [lucene]

2023-12-26 Thread via GitHub
dungba88 commented on PR #12980: URL: https://github.com/apache/lucene/pull/12980#issuecomment-1869917203 Fixed the above unclosed issue by moving `openInput` and `createOutput` to try-catch block -- This is an automated message from the Apache Git Service. To respond to the message, plea

Re: [PR] Make FSTPostingFormat to build FST off-heap [lucene]

2023-12-26 Thread via GitHub
dungba88 commented on PR #12980: URL: https://github.com/apache/lucene/pull/12980#issuecomment-1869867635 There is only 1 failed test left: TestFSTPostingFormat.testRandomException ``` > Caused by: > java.lang.RuntimeException: unclosed IndexOutput: _s_FST5

Re: [PR] Move group-varint encoding/decoding logic to DataOutput/DataInput [lucene]

2023-12-26 Thread via GitHub
uschindler commented on PR #12841: URL: https://github.com/apache/lucene/pull/12841#issuecomment-1869642731 Hi, Thanks for the measurements, @easyice. So basically, we can backport the commit without any modifications. I can do this and move change entries afterwards. No PR is

Re: [PR] Move group-varint encoding/decoding logic to DataOutput/DataInput [lucene]

2023-12-26 Thread via GitHub
easyice commented on PR #12841: URL: https://github.com/apache/lucene/pull/12841#issuecomment-1869575788 There seems to be a speedup on [prefix queries](http://people.apache.org/~mikemccand/lucenebench/Prefix3.html) in nightly benchmarks. For reference, here is the benchmark in branc

Re: [I] FSDirectory stuck at open(Path path) method when ran from .jar file [lucene]

2023-12-26 Thread via GitHub
uschindler commented on issue #12968: URL: https://github.com/apache/lucene/issues/12968#issuecomment-1869530344 > creating a "org.apache.lucene.codecs.PostingsFormat" file into the "META-INF/services" directory solved the issue. You shouldn't manually create those files. Use https:/

Re: [I] [BUG] FSDirectory stuck at open(Path path) method when ran from .jar file [lucene]

2023-12-26 Thread via GitHub
setokk commented on issue #12968: URL: https://github.com/apache/lucene/issues/12968#issuecomment-1869471695 Thanks! Adding the line "Multi-Release: true" to the MANIFEST.MF file and creating a "org.apache.lucene.codecs.PostingsFormat" file into the "META-INF/services" directory solved the

Re: [I] [BUG] FSDirectory stuck at open(Path path) method when ran from .jar file [lucene]

2023-12-26 Thread via GitHub
uschindler commented on issue #12968: URL: https://github.com/apache/lucene/issues/12968#issuecomment-1869469388 Most likely you see this issue due to broken Maven tooling: https://issues.apache.org/jira/browse/MSHADE-385 -- This is an automated message from the Apache Git Service. To res

Re: [I] [BUG] FSDirectory stuck at open(Path path) method when ran from .jar file [lucene]

2023-12-26 Thread via GitHub
uschindler commented on issue #12968: URL: https://github.com/apache/lucene/issues/12968#issuecomment-1869432712 The problem is that your packaging as JAR does not preserve all required files. See here for instructions: #12307 It looks like you suppress some exceptions. The pro

Re: [I] [BUG] FSDirectory stuck at open(Path path) method when ran from .jar file [lucene]

2023-12-26 Thread via GitHub
uschindler closed issue #12968: [BUG] FSDirectory stuck at open(Path path) method when ran from .jar file URL: https://github.com/apache/lucene/issues/12968 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to

[PR] lazily write the FST padding byte [lucene]

2023-12-26 Thread via GitHub
dungba88 opened a new pull request, #12981: URL: https://github.com/apache/lucene/pull/12981 ### Description Lazily write the FST padding byte, so that in case the FST is empty (no accepted nodes) nothing will be written. This is important for off-heap writing, as we don't want to ad

Re: [PR] Make FSTPostingFormat to build FST off-heap [lucene]

2023-12-26 Thread via GitHub
dungba88 commented on PR #12980: URL: https://github.com/apache/lucene/pull/12980#issuecomment-1869385378 I found another quite tricky issue: If we write the FST directly to the IndexOutput, there might be a chance that there's no term accepted by the FST, in that case we still write